If AI removes the labor constraint on high-skill work, what happens to the advantage of elite firms?

Ken Griffin recently said he went home one Friday “fairly depressed” after watching AI agents at Citadel complete work in days that previously took teams of finance PhDs months. He specifically said this was not low-skill work being automated, but “extraordinarily high skilled jobs.” From an economic perspective, if frontier AI meaningfully reduces the labor constraint on highly specialized work, does that compress the advantage of elite firms? For example, if someone straight out of college can run quant-level analysis or research workflows that previously required large teams of elite talent, what happens to the value of human capital, institutional scale, and information asymmetry over time? Smart people no longer need to work at Citadel?